What Is a CDP and How Does It Differ From Other Customer Data Tools?

Customer data platforms (CDPs) are picking up steam: Gartner named them one of six marketing technologies to watch in 2018 and the category is expected to reach $1 billion in revenue by 2019.

Customer data platforms (CDPs) are picking up steam: Gartner named them one of six marketing technologies to watch in 2018 and the category is expected to reach $1 billion in revenue by 2019.

Publishers have actually been early adopters when it comes to CDPs, particularly because CDPs enable them to make greater use of first-party data (which publishers have a lot of) to drive content recommendations and individual-level ad targeting. Now more marketers are taking an interest in CDPs, which is expected to grow as the industry moves further toward permission-based marketing and is focused on providing more unified cross-platform experiences.

One of the first questions marketers may have about CDPs is what exactly they are and how they differ from other customer databases, notably data management platforms (DMPs). Let’s try to reduce that confusion.

Definitions first. The Customer Data Platform Institute, which I founded in 2016, defines a CDP as a “marketer-managed system that creates a persistent, unified customer database that is accessible to other systems”. That’s open to lot of interpretation. In more concrete terms, what CDPs do is:

Gather All Customer Data in One Place

That’s the “persistent, unified customer database” part of the definition. It distinguishes CDPs from systems that store data without linking items related to the same customer (like many “data lakes”), that gather customer data on the fly (like many behavior-based personalization products), or that hold just an anonymous cookie ID plus audience attributes (like a Data Management Platform). The CDP stores all details down to the level of web pageviews, email clicks, purchases, payments, and content selections. And it stores personal identifiers so data can be assembled into a single customer view and so messages can be directed at known individuals. One of the reasons that publishers have eagerly adopted CDPs is that detailed, personally identified data is exactly what they needed for content recommendations and individual-level ad targeting, two key profit drivers.

Allow Other Systems to Use the Assembled Data

This distinguishes CDPs from systems that build a single customer view for their own use, including many recommendation engines, personalization tools, and email platforms. Access can range from direct queries against a database table, to periodic file exports, to calls to an Application Program Interface (API). Sometimes the external systems read raw CDP data and sometimes they read data that’s pre-summarized and reformatted. Whatever the technical details, the key point is the CDP data isn’t locked away where no other system can use it.

Make Data Relatively Easy to Manage

“Relatively” is slippery weasel of a word, but bear with me. If anyone tells you they can build a persistent, unified customer database without any technical skills, just back away slowly, turn, and run for the nearest exit because they’re either dishonest or deluded. CDPs reduce the technical workload by using innovations such as “schema-less” data stores that ingest data without pre-defining its contents. They also contain prebuilt components, such as identity matching processes, that save time over building similar software from scratch. The net result is that CDPs can build and update customer databases much faster and with many fewer technical work-hours than conventional systems. This in turn means that marketers can request changes more freely because the cost of each change is so low. It’s hard to quantify the difference without looking at specific situations, but it’s quite realistic to expect changes that took weeks in a conventional system to take days or just hours with a CDP.

You’re probably skeptical of these claims. You should be. You’ll need some quality time with a CDP vendor to be convinced. But just suspend disbelief for a moment to imagine what you could do with a CDP if it actually performed as promised.

CDPs support two main applications. The first is analytics: by bringing together all customer data, cleaning it, creating a unified customer view, and making the result available to other systems, the CDP lets your analysts spend less time on wrangling data and more time creating value. This isn’t a little benefit: analysts spend as much as 80% of their time finding and preparing data.

The second application is execution support. CDPs can feed data to order management systems, websites, email engines, call centers, ad networks, and even data management platforms to support segmentation, personalization, campaign management, and content recommendations.

But there’s more. Quite a few CDP systems include their own analytical and execution applications. This further increases deployment speed and value provided.

In short, CDPs are different from your existing CRMs, data lakes, DMPs, and old-style marketing databases. They combine detailed data, low cost, and unprecedented flexibility to create value throughout a publishing operation. It may sound too good to be true. But it’s surely worth some time to find out.